diff mbox series

[for-2.11] pc: fix crash on attempted cpu unplug

Message ID 1511188453-248734-1-git-send-email-imammedo@redhat.com
State New
Headers show
Series [for-2.11] pc: fix crash on attempted cpu unplug | expand

Commit Message

Igor Mammedov Nov. 20, 2017, 2:34 p.m. UTC
when qemu is started with '-no-acpi' CLI option, an attempt
to unplug a CPU using device_del results in null pointer
dereference at:

  #0 object_get_class
  #1 pc_machine_device_unplug_request_cb
  #2 qmp_marshal_device_del

which is caused by pcms->acpi_dev == NULL due to ACPI support
being disabled.

Considering that ACPI support is necessary for unplug to work,
check that it's enabled and fail unplug request gracefully
if no acpi device were found.

Signed-off-by: Igor Mammedov <imammedo@redhat.com>
---
 hw/i386/pc.c | 5 +++++
 1 file changed, 5 insertions(+)

Comments

Eduardo Habkost Nov. 20, 2017, 2:44 p.m. UTC | #1
On Mon, Nov 20, 2017 at 03:34:13PM +0100, Igor Mammedov wrote:
> when qemu is started with '-no-acpi' CLI option, an attempt
> to unplug a CPU using device_del results in null pointer
> dereference at:
> 
>   #0 object_get_class
>   #1 pc_machine_device_unplug_request_cb
>   #2 qmp_marshal_device_del
> 
> which is caused by pcms->acpi_dev == NULL due to ACPI support
> being disabled.
> 
> Considering that ACPI support is necessary for unplug to work,
> check that it's enabled and fail unplug request gracefully
> if no acpi device were found.
> 
> Signed-off-by: Igor Mammedov <imammedo@redhat.com>

Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>

But I have one small suggestion:

> ---
>  hw/i386/pc.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> index c3afe5b..d80cec3 100644
> --- a/hw/i386/pc.c
> +++ b/hw/i386/pc.c
> @@ -1842,6 +1842,11 @@ static void pc_cpu_unplug_request_cb(HotplugHandler *hotplug_dev,
>      X86CPU *cpu = X86_CPU(dev);
>      PCMachineState *pcms = PC_MACHINE(hotplug_dev);
>  
> +    if (!pcms->acpi_dev) {
> +        error_setg(&local_err, "not supported without acpi");

I suggest "CPU hot unplug not supported without ACPI" instead.

> +        goto out;
> +    }
> +
>      pc_find_cpu_slot(MACHINE(pcms), cpu->apic_id, &idx);
>      assert(idx != -1);
>      if (idx == 0) {
> -- 
> 2.7.4
>
Igor Mammedov Nov. 20, 2017, 2:59 p.m. UTC | #2
On Mon, 20 Nov 2017 12:44:54 -0200
Eduardo Habkost <ehabkost@redhat.com> wrote:

> On Mon, Nov 20, 2017 at 03:34:13PM +0100, Igor Mammedov wrote:
> > when qemu is started with '-no-acpi' CLI option, an attempt
> > to unplug a CPU using device_del results in null pointer
> > dereference at:
> > 
> >   #0 object_get_class
> >   #1 pc_machine_device_unplug_request_cb
> >   #2 qmp_marshal_device_del
> > 
> > which is caused by pcms->acpi_dev == NULL due to ACPI support
> > being disabled.
> > 
> > Considering that ACPI support is necessary for unplug to work,
> > check that it's enabled and fail unplug request gracefully
> > if no acpi device were found.
> > 
> > Signed-off-by: Igor Mammedov <imammedo@redhat.com>  
> 
> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
> 
> But I have one small suggestion:
> 
> > ---
> >  hw/i386/pc.c | 5 +++++
> >  1 file changed, 5 insertions(+)
> > 
> > diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> > index c3afe5b..d80cec3 100644
> > --- a/hw/i386/pc.c
> > +++ b/hw/i386/pc.c
> > @@ -1842,6 +1842,11 @@ static void pc_cpu_unplug_request_cb(HotplugHandler *hotplug_dev,
> >      X86CPU *cpu = X86_CPU(dev);
> >      PCMachineState *pcms = PC_MACHINE(hotplug_dev);
> >  
> > +    if (!pcms->acpi_dev) {
> > +        error_setg(&local_err, "not supported without acpi");  
> 
> I suggest "CPU hot unplug not supported without ACPI" instead.
I've thought about it but considering error is issued in context of
device_del command, I've opted for more concise variant.

Would you still like me to respin patch with your suggestion?

> 
> > +        goto out;
> > +    }
> > +
> >      pc_find_cpu_slot(MACHINE(pcms), cpu->apic_id, &idx);
> >      assert(idx != -1);
> >      if (idx == 0) {
> > -- 
> > 2.7.4
> >   
>
Eduardo Habkost Nov. 20, 2017, 5:05 p.m. UTC | #3
On Mon, Nov 20, 2017 at 03:59:51PM +0100, Igor Mammedov wrote:
> On Mon, 20 Nov 2017 12:44:54 -0200
> Eduardo Habkost <ehabkost@redhat.com> wrote:
> 
> > On Mon, Nov 20, 2017 at 03:34:13PM +0100, Igor Mammedov wrote:
> > > when qemu is started with '-no-acpi' CLI option, an attempt
> > > to unplug a CPU using device_del results in null pointer
> > > dereference at:
> > > 
> > >   #0 object_get_class
> > >   #1 pc_machine_device_unplug_request_cb
> > >   #2 qmp_marshal_device_del
> > > 
> > > which is caused by pcms->acpi_dev == NULL due to ACPI support
> > > being disabled.
> > > 
> > > Considering that ACPI support is necessary for unplug to work,
> > > check that it's enabled and fail unplug request gracefully
> > > if no acpi device were found.
> > > 
> > > Signed-off-by: Igor Mammedov <imammedo@redhat.com>  
> > 
> > Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
> > 
> > But I have one small suggestion:
> > 
> > > ---
> > >  hw/i386/pc.c | 5 +++++
> > >  1 file changed, 5 insertions(+)
> > > 
> > > diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> > > index c3afe5b..d80cec3 100644
> > > --- a/hw/i386/pc.c
> > > +++ b/hw/i386/pc.c
> > > @@ -1842,6 +1842,11 @@ static void pc_cpu_unplug_request_cb(HotplugHandler *hotplug_dev,
> > >      X86CPU *cpu = X86_CPU(dev);
> > >      PCMachineState *pcms = PC_MACHINE(hotplug_dev);
> > >  
> > > +    if (!pcms->acpi_dev) {
> > > +        error_setg(&local_err, "not supported without acpi");  
> > 
> > I suggest "CPU hot unplug not supported without ACPI" instead.
> I've thought about it but considering error is issued in context of
> device_del command, I've opted for more concise variant.
> 
> Would you still like me to respin patch with your suggestion?

I'd prefer to, so the error message would make sense even if out
of context.  But you still have my Reviewed-by in case Michael
wants to apply this version.


> 
> > 
> > > +        goto out;
> > > +    }
> > > +
> > >      pc_find_cpu_slot(MACHINE(pcms), cpu->apic_id, &idx);
> > >      assert(idx != -1);
> > >      if (idx == 0) {
> > > -- 
> > > 2.7.4
> > >   
> > 
>
Paolo Bonzini Nov. 28, 2017, 11:28 a.m. UTC | #4
On 20/11/2017 18:05, Eduardo Habkost wrote:
> On Mon, Nov 20, 2017 at 03:59:51PM +0100, Igor Mammedov wrote:
>> On Mon, 20 Nov 2017 12:44:54 -0200
>> Eduardo Habkost <ehabkost@redhat.com> wrote:
>>
>>> On Mon, Nov 20, 2017 at 03:34:13PM +0100, Igor Mammedov wrote:
>>>> when qemu is started with '-no-acpi' CLI option, an attempt
>>>> to unplug a CPU using device_del results in null pointer
>>>> dereference at:
>>>>
>>>>   #0 object_get_class
>>>>   #1 pc_machine_device_unplug_request_cb
>>>>   #2 qmp_marshal_device_del
>>>>
>>>> which is caused by pcms->acpi_dev == NULL due to ACPI support
>>>> being disabled.
>>>>
>>>> Considering that ACPI support is necessary for unplug to work,
>>>> check that it's enabled and fail unplug request gracefully
>>>> if no acpi device were found.
>>>>
>>>> Signed-off-by: Igor Mammedov <imammedo@redhat.com>  
>>>
>>> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
>>>
>>> But I have one small suggestion:
>>>
>>>> ---
>>>>  hw/i386/pc.c | 5 +++++
>>>>  1 file changed, 5 insertions(+)
>>>>
>>>> diff --git a/hw/i386/pc.c b/hw/i386/pc.c
>>>> index c3afe5b..d80cec3 100644
>>>> --- a/hw/i386/pc.c
>>>> +++ b/hw/i386/pc.c
>>>> @@ -1842,6 +1842,11 @@ static void pc_cpu_unplug_request_cb(HotplugHandler *hotplug_dev,
>>>>      X86CPU *cpu = X86_CPU(dev);
>>>>      PCMachineState *pcms = PC_MACHINE(hotplug_dev);
>>>>  
>>>> +    if (!pcms->acpi_dev) {
>>>> +        error_setg(&local_err, "not supported without acpi");  
>>>
>>> I suggest "CPU hot unplug not supported without ACPI" instead.
>> I've thought about it but considering error is issued in context of
>> device_del command, I've opted for more concise variant.
>>
>> Would you still like me to respin patch with your suggestion?
> 
> I'd prefer to, so the error message would make sense even if out
> of context.  But you still have my Reviewed-by in case Michael
> wants to apply this version.

I made the change and queued the patch.

Paolo
Igor Mammedov Nov. 28, 2017, 1:23 p.m. UTC | #5
On Tue, 28 Nov 2017 12:28:43 +0100
Paolo Bonzini <pbonzini@redhat.com> wrote:

> On 20/11/2017 18:05, Eduardo Habkost wrote:
> > On Mon, Nov 20, 2017 at 03:59:51PM +0100, Igor Mammedov wrote:  
> >> On Mon, 20 Nov 2017 12:44:54 -0200
> >> Eduardo Habkost <ehabkost@redhat.com> wrote:
> >>  
> >>> On Mon, Nov 20, 2017 at 03:34:13PM +0100, Igor Mammedov wrote:  
> >>>> when qemu is started with '-no-acpi' CLI option, an attempt
> >>>> to unplug a CPU using device_del results in null pointer
> >>>> dereference at:
> >>>>
> >>>>   #0 object_get_class
> >>>>   #1 pc_machine_device_unplug_request_cb
> >>>>   #2 qmp_marshal_device_del
> >>>>
> >>>> which is caused by pcms->acpi_dev == NULL due to ACPI support
> >>>> being disabled.
> >>>>
> >>>> Considering that ACPI support is necessary for unplug to work,
> >>>> check that it's enabled and fail unplug request gracefully
> >>>> if no acpi device were found.
> >>>>
> >>>> Signed-off-by: Igor Mammedov <imammedo@redhat.com>    
> >>>
> >>> Reviewed-by: Eduardo Habkost <ehabkost@redhat.com>
> >>>
> >>> But I have one small suggestion:
> >>>  
> >>>> ---
> >>>>  hw/i386/pc.c | 5 +++++
> >>>>  1 file changed, 5 insertions(+)
> >>>>
> >>>> diff --git a/hw/i386/pc.c b/hw/i386/pc.c
> >>>> index c3afe5b..d80cec3 100644
> >>>> --- a/hw/i386/pc.c
> >>>> +++ b/hw/i386/pc.c
> >>>> @@ -1842,6 +1842,11 @@ static void pc_cpu_unplug_request_cb(HotplugHandler *hotplug_dev,
> >>>>      X86CPU *cpu = X86_CPU(dev);
> >>>>      PCMachineState *pcms = PC_MACHINE(hotplug_dev);
> >>>>  
> >>>> +    if (!pcms->acpi_dev) {
> >>>> +        error_setg(&local_err, "not supported without acpi");    
> >>>
> >>> I suggest "CPU hot unplug not supported without ACPI" instead.  
> >> I've thought about it but considering error is issued in context of
> >> device_del command, I've opted for more concise variant.
> >>
> >> Would you still like me to respin patch with your suggestion?  
> > 
> > I'd prefer to, so the error message would make sense even if out
> > of context.  But you still have my Reviewed-by in case Michael
> > wants to apply this version.  
> 
> I made the change and queued the patch.
I've thought Michael (CCed) queued v2
but he hasn't sent a pull req yet.

> Paolo
diff mbox series

Patch

diff --git a/hw/i386/pc.c b/hw/i386/pc.c
index c3afe5b..d80cec3 100644
--- a/hw/i386/pc.c
+++ b/hw/i386/pc.c
@@ -1842,6 +1842,11 @@  static void pc_cpu_unplug_request_cb(HotplugHandler *hotplug_dev,
     X86CPU *cpu = X86_CPU(dev);
     PCMachineState *pcms = PC_MACHINE(hotplug_dev);
 
+    if (!pcms->acpi_dev) {
+        error_setg(&local_err, "not supported without acpi");
+        goto out;
+    }
+
     pc_find_cpu_slot(MACHINE(pcms), cpu->apic_id, &idx);
     assert(idx != -1);
     if (idx == 0) {